Tobacco Documents Online

c Persons who died between 1975 anii 1979 could not be accounted for in these estimates. Because smokers experience a higher death rate than nonsmokers (25), the figures in the table somewhat overestimate the percentage o6 lifetime nonsmokers, mainly in the older age groups. Because of the higber prevalence of smoking among males, this overestimation is probably greater for them than for-females„ and it thus may exert a stronger upward bias im subsequent calculations of male than of female nonsmokers' life expectancv. A simulation of this biasing effect~ifi we assume a high smoker-to-nonsmoker mortality.ratio for all cohorts-yields an upward bias for males of less than one-half year in life /ucpectancy. We con- sidered this difference too small to require adjust- ment of the raw data for the life table, but we have taken it into account in interpreting the life table results. Population estimates for 1973, provided by the ! Division of Health Statistics and Research, Pennsyl- vania DepartmenG of Health, were multiplied by 3 because the mortality data covered 33 years, and' the results were then multiplied by the proportions of nonsmoking men and women in the top table on page 346. These computations provided the denom- inators needed for calculating the mortality rates for each age-sex cohort, the numerators being the non- smoker deaths recorded in the Northwesterm Penn- sylvania Study on Smoking and Healthl The num- ber of deaths,in each age group in column 2 of the life table was divided by the total number of persons in the living population in column 3 to establish the age-specific mortality rate. in column 4. These rates were multiplied'by the Chiang constants in column 5 to produce the adjusted probability function (Q,) in column 6 (26). The life table computations im col- umns 7 to 10 . were done according to the standard demographic procedures used by Shryock and Siegeli (27): /, = number living at the beginning of the age interval; d, = number dying,duringthe age interval; . L, = stationary population in the age interval; T. = stationary population in the designated age in- terval plus all subsequent age intervals; and e, = life expectancy at entry to each age eategory (reported in column 11 ot the life tables): - ~ The table shows that values for life expectancy are similar for both male andd female nonsmokers in all age groups above 30 years. Standard errors for life expcctancy at age 30, calculated according to the method of Kcyfitz (28), are 0.55 years for men and 0.30 years for women.. The standard errors de- cline for each older age group. The life expectancy values for men and women overlap the standard errors in each age group, but the male expeetancies are on the whole slightly higher, possibly because of the bias already discussed in connection with esti- mating the base percentage of nonsmokers. Discussion The table shows a consistent pattern of similarity in the life expectancy for all 5-year cohorts of non- smoking mem and women over -age 30 in Erie County. The nearly identical life expectancy that' is foundwhen traumatic deaths (fatal accidents, homi- cides, and suicides) are removedand better pro- cedures for classifying nonsmokers are instituted corroborates our hypothesis that differences in smok- ing habits are responsible for observed male-female - longevity differences. The result supports Preston's conclusion that virtually all the increase in the dif- ference between male an& female longevity since 1930 is attributable to the effects of cigarette smok- I; ing. The data from our study, combined with more I than four decades of research showing the destruc- tive force of cigarette smoke and the fact that men i have a greater number of smoking years than women (25),, provides ample evidence of the impact that smoking has on the MFLD. Our results also agree withthose ot Casey and Casey (19) and Miller (20). Although all studies in which the MFLD hasbeen investigated have revealed, a substantial detrimental impact of cigarette.smoking, several of the investiga- tors have reported residual MFLD not accounted for by smoking. There are two likely explanations for this residual difference. First, according to the Vital Statistics of the United States, traumatic deaths due to accidents,, suicides, and homicides occur in the greatest numbers among men in the lower age brack- ets-ages 20 through 55-and these relatively early deaths produce a disproportionate impact on lon- gevity statistics. In our study, this effect was elim- inated. In any new research on the MFLD or in re- examination of earlier studies; investigators will need to take traumatic deaths into account. 0373470 Second, a review of the methods used in other studies shows ambiguities in distinguishing between { nonsmokers and former smokers. An indeterminate i number of interviewees, when answeringg questions. I abouetheir own smoking habits or those of their rela- tives, initially classify themselves or their refatives as `I nonsmokers even though they may have smoked I the past. This inaccuracy is particularly common among more distant relatives and among younger Jvqduvue .f1[]. Yol. N, Me.f 8r

/ two sets. of data must be obtained instead of one, and the possibility exists, that the representation will not be as precise. Using the two-samRle,, cross-sectional technique, we combined previously collected retrospective data (21) om the lifetime smoking habits of adult men an& women who had~ died in Erie County, Pa., in the yeart 1972-74 (data derived from interviews with close relatives of the deceased) with new lifetime survey data on the smoking habits of men and women who. lived in Erie County 1972-74 (data we obtained by reverse projection from a 1979 sur- vey). In both surveys, telephone interviews were used because we believed that the data obtained in this way would be more accurate than data from self-administered qpestionnaires. Identical items were used in the interviews in botb surveys. Roget and: Reid (22) have shown. that data on decedents' smoking habits whi& are carefully collected from survivors are as reliable as data obtained on living popu l ations. Study Population Eric is, a medium-sized urban area in the north- eastern United States. According to the 1970 U~S. Census, the population of Erie County was 263,654. The city of Erie was,the third largest municipality in Pennsylvania and had a population of 129,341. The county has had a history of low migratory rates. According to records provided by the Pennsylvania Department of Health, Division of Health Statistics and' Research, the total in and out migration in the last 20 years has been approximately 7 percent. In the years,1972-74, Erie County had an annual death rate of 9.9 per 1,000, which is typical for a north. eastern Standard Metropolitan Statistical Area. The male-female difference in lift expectancy for persons in Erie County age 30 and older was approximately 6 years in favor of women, a difference that closely approximates the national figure. Data on the decedent population. In 11973„ Miller (21) began the Northwestern Pennsylvania Study on Smoking and Health to determine retrospectively the smoking habits of deceased residents of Erie County based on telephone interviews with the de- cedents' relatives. Death notices for the ytar,s 19727 74 (their publication in a local newspaper is a usual requirement) provided: the names of nearly even, person who had died in Erie County in these years along with the names of their closest surviving next of kio. Im the study, telephone numbers oG up to. three surviving relatives were identified for each C death notice, if possible. No telephone number could be obtained, however, for the survivors of 15 per- cent of the decedents because, for example, they had' no surviving relatives in Eric County, telephone list- ings were not available, or the decedent was a tran- sient. In order to focus on the impact of smoking, all deaths caused by accident, homicide, or suicide and all decedents under age 30 were eliminated from the analysis. Wlten telephoning, the interviewers briefly ex- plained the purposes of the. study to the relatives they contacted and solicited their cooperation. In- formation was collected on the exact cause of death andthe decedent's age, occupation, and smoking his- totryry. The interviewers recorded any detailed com- ments by the respondent and coded the data. In the National. Mortality Survey (17.18) and the earlier survey of lung cancer mortality upon which it was modeled (16), survivors' reports on dece- dents' smoking habits were compared with the dece- dents' medical records and the decedents' own reports before death. The data provided by the sur- vivors were found to be satisfactory in~ respect to recent smoking status; agreement was nearl• perfect. li Hbwever, the distinction between lifetime non I smokers and long-time former smokers was much less reliable. Some survivors did not report the de- cedent's former smoking when responding on the questionnaire to the single item that I distinguished between lifetime nonsmokers and for- mer smokers. In our telephone survey, therefore, particular attention was paid to probing deeper into ; an initial response about prior use of cigarettes if the decedent was reported as not having smoked. Although the interviewers called the telephone numbers of survivors five times or more if neces- san•, no telephone contact could be made in ap- proximately 10 percent of the cases. However, of I the relatives.contacted, 95 percent provided the in- formation needed for the study. Of the 6,930 per- sons aged 30 and older who,haddied in Eric County Iin the vears 1972-74, usable interviews were ob- tained for 63 percent, or 4,394 decedents. For the purposes of this report. only the lifetime non-. smokces-2.195 persons who had smoked less than 20 packs of cigarettes during theit lifetime-were considered for analysis. U3734699~: .. Data on the living populotion. To determine the smoking habits of the living population of Erie County for the years 1972-74, a 2 percentt randomm sample of household telephone numberas was. taken from Erie County telephone directories for these years. About 88 percent of the households in the JvIrJUYu.l 1993. Yn6.91. Me. 4 7!!

. related' accidents and', Alcohol, Drug Abuse, and Mental Health Services Block Grants. that require States to set aside funds for alcoholism prevention progams. ) The students.were encouraged by these and other Federal efforts to alert the country to the dangers of drunk driving. As one teenager put it, "National recognition like this will makee it easier for us to get our communities moving." . .,, The conference closed on an exuberant note as Dr. William Mayer, administrator of the Alcohol, Drug, Abuse, and Mental Health Administration and acting director`of the National Institute om Alcohol Abuse and Afcoholisml awarded certificates of achievement to the smdents who had presented th ' eight model projects throughout the weekend. "Your presence here places you~ among e new leaders of the country," Jvfayer told' th students. "From what I've seen this weekend, th uture of our nation is in good hands." .., Repeating an earlier pledge b ecretary Heckler, Mayer announced that the paptment of Health and' Human Services inten to hold similar confer- ences annually to mo ' future generations of students against d[un riving. uppflK Nearly 5,(25,000 in private contributions from 10 organiz ons and foundations helped sponsor this year' conference. Coordinators were James Kemper, Jr., chairman of the board of Kemper Insurance Group, and Davis Taylor, member of the Board of The Life Expectancy of Nonsmoking Men and Women G. H. MILLER„ PhD DEAN R. GERSTEIN, PhD Dr. Miller is dhmrn of Studies on Smokieg (SOS).,, a research program in Edinboro,. Pa., that includes smoking cessation clinics. Dr. Gerstein is studgdirector,"CommissioK on Behavioral and Social Sciences and Education, National Research CAUncil,. Washington,. D.C. Parts of the work re- ported here were supponedby a grant fromm the National Institute on Drug Abuse (Contract No. 271-76-331) tothe Committee.on Substance Abuse and Habitual Behavior. Ne• tional Research Counoil,, and by the Eric Chapter ofthe ~ American Cancer Society, the. Northwestern Pennsylvania T raiiapottation Secretary Elizabeth DoIe: ~Drunk driving E~:. -' =er-- s` homindie~,;dnd th ' s-•:a.u$ers,from it =:;gerterdtion . , .,., -,:brivera: l re;,respo W ain Directors, Boston Globe Foundation, Inc. Both men received the Secretary's V olunteer Award' for stimtr lating privatee sector support of the conference. Other contributors included: Allstate Foundation; Dow Jones and Company, Inc.; GEICO. Philan- thropic Foundation; Philip L Graham Fund (The Washington Post); Knight-Ridder Newspapers, Inc.; Lee Enterprises; Liberty Mutual Insurance Com- pany; and The New York Times. References ......................... 1. Department of Health, Educationt andWelfare: Hr:ahby people-the Surgeon General's report on bealth promo- tion and disease: prevention. DHEW (PH5) Publication No. 79-55071. U.S. Government Printing Office, Wash• ington,D.C., 1979, p. 44. 2.. Malin, H., Trumble. I., Kaelbcq C.,. and, Lubran, B.: Alcoho1-related highway fatalities among young drivers -United. States. MMWR 31: G41-6d4i Dec. 10, 1982. Lung Association, and the Heart Assoeiation of Prie County. The article is based on a presentation by Dr. Miller to the Slh World Conference on Smoking and Health„ in Win- nipeg. Manitoba, Canada, luly. 14, 1983.Tearsheet requests to Dr. G. H. Miller.. Director,. Studies on Smoking. (SOS), 711-2Cireleville Rd., State College, Pa. 16801. SYNOPSIS .............. 03734E~,9'7 . The pronounced difference in life expectancy be- tween men and women in the United States and other industrialized countries has been attributed to a varietyy of causes, among them, diBerential rates oJcigarette smoking. A.study was underrakentoetiminate the confounding factors of imprecision in the taking of smoking histories and exaggeration of . JvtM.oust SMa. vet. q, Me. 4 . 043

earl ' ~ traumatic deaths: in life expectancy calcu- lations. Survev data weree collected on the lifecime smok- ing habits of adrtYts in Erie County, Pa., as of 1972- 74. In the survey interviews, careful distinctions were made between respondents who had formerly smoked andTespondents who had never smoked. The survey data were combined with data collected from surviving relatives about the smoking habits of peo- ple who had died in Eric County during the years 1972-74. After deathsattributabfe to traumatic THE nIFFERENCE IN LIFE F.XPECTANCY between U.S, males and females is substantiali and' has been increasing for many years. A longevity difference of 2 years in favor of women in the earlypart of the 20th century widened to a difference of 8 years in women's favor in 1979' (1). A number of theories have been proposed to explain this difference., Fisher (2), Berkson (3), Montagu (4), and others have attributed the male-female longevity difference (MFLD) to genetic differences. Lond'on and asso- ciates (5) and Stamler and associates (6) proposed that higher estrogen levels provided women with dif- ferential' protection from cardiovascular diseases, thus enabling them to live longer than men. Rosen- man and' associates (7), Russek (8), Jenkins (9,10), Haves and Feinleib (11), and Waldron (12) have proposed: that the stress associated with the type A behavior pattern that is so prevalent among males contributes significantly to increased cardiovascular disease. The impact of smoking on mortality has been the subject of extensive scientific investigation, but few of the studies have dealt directlv with the male- female longevity difference. Preston (13),.analyzing mortality changes in 16 countries, concluded that the internatdonalincreasein MFLD from1930. to 1963 wasdue largely to cigarette smoking. Retherford (14), using datafrom the American Cancer So- ciety's sample of 1' million volunteers (15), esti- mated that less than ttal8 (47 percent) o6 the U.S MFLD in 1962 was due to cigarette smoking. Sur- veys by Haenszel and: Associates (1b), Godlev (17), and Enstrom and Godley (18) indicated sotre differences in the MFLD in the nonsmoking segment of their samples. Casey and, Casey (19) in Ireland and, Miller (20) in the United States investigated data from certain rural communities where cigarette smoking was virtually nonexistent and found no female longevity advantage. causes (accidents, suicides, and homicides) were re- moved, lite tables were calculated (or male and fe- male nonsmokers over age 30. The resulting lifees- pectancy (tguresfor nonsmoking, men and women of parallel age were virtually identical. Thus, dfjferen- tial rates of cigarette smoking are apparently the overwhelming cause for the male-female longevity di.Uerence. Actuarial tables: should be divided by smokiitg behavior to reflect this finding. The results of the study suggest that the presentlongevity.diBer- ence between men and women will disappear. The mortality data in the studies reviewed gen- erally included all causes of death or the deaths from lung cancer. We hypothesized that the differences among results related to male-female longevity might be due to varying incidences of deaths from trau- matic causes, such as fatal accidents and suicides, and to differences in the methods used to identify and classify study participants as.nonsmokers. There- .fore we undertook to test the hypothesis that no male-female longevity differences would be found if the effects of fatal injury were removed and par- ticular care was taken to exclude former smokers from the nonsmoker category. Design ot the Study The statistical method used in the study was, a two-sample, cross-sectional analysis devised by Haenszel and associates (16). The advantage of this technique is that by combining data from a deceased population (numerator data) with data from a liv- ing population (denominator data), good estimates of mortality ratios and life tables can be obtained in a much shorter time than life tables generated by prospective studies. In this kind'of study, however, 031734698 _dT < y'':a 'We hypothesrzed that the di$erence ! . . . J.;.among results,relcfed to rnale j•emale.'.t 3''; 7ongevity migRt be due to varytng ' r~ l ~ ° incidences of cfeafhs1roni traumattc_ . , ; catues, such as fatal accidents end .. ;, stricides, and_to differences in the ;; -,.T:t:: ; methods used to identify and classifv. ~ sruay Parttcipantsasnonsrtlokers,r;-;L4t`, W ~Pvb11e NPPIM R.prIP Y 1 'n

Erie 5tandard Metropolitan Statistical Area uad tele- phones available (23,2d).,,The names listed and the telephonenumbers.were compared with those in'thc .1979 direetory,. A eurrent hsting was not' obtained forapproximatJy ]4.pereent of the 1972-)4 sample because the persons had moved from the arez, had subscquenth' obtained an unlisted numbcr„ or had discontinued telephoneserviee. , . The items used in telephone interviews with the living population were identical to those used: in interviewing the relatives off decedents exeept that jnappliczble items such as ••Cause of.f death" were omitted. Information on each household mcmber 30 years of age or older was obtainedi.The telephone in- terviewjngtlegan in April 1979 and was tompleted._~in ]uly 1979.Informatfon wasobtajned on a to:alof 3,916 residents of Erie County.a 96 percent re- sponse sponse ra¢...This information was themanalyzed to .I det rmme_whcther in the year of the directorp list. ing from which theresponoenUS name had been - drawn (the "base year"), the respondent had been _ a currenteiFaretl^ smoi<er,l a former smoker (at'.. leas 1 year had elapsed since quitttng),, or a non- -C smoker (less tham20 packssmokcd during ba or ller lifetlme). The. National Researeh Counoil's Com- mittee.oniSubstance Abuse and Habitual Behavi0r assisted with this segment of the study. . r . Analysis of Lile Tables -, _ . . The percentages ob nonsmoking meniandwomen by age group, as retrospectivelyy esti i_in the 1972-74 population of Erie. County, were as fo}- )ows; ....._ .. _ .- . ... .. . .~.. - . - S _vear . . _- No.umeking NnnlmokfnF-eoho r , . _ . mee .;,.., women '. z. ..... -. .... . 31,7 59:9 '35-09 ....:....:.:......~..,.... 31.k .. . 56,7 46.4a'..~..:....:d,..~...:r.: ' 2]]. .- ..'. 59.6 <5_49.....:.:.I..'.i..:......r, 24.4 6I.0 . 50-54'~ ....:.................... 30.9. .. n' 59.8 55-59'~ ......... ........ 26.7 ._ 6<J ~6D_64 . ' 21,] ~62d ~ 61-69 ..,, ~263 ~ '622 70.74 ,,. 30.6 T'2 75-79 ... . 35, 82< 80-84 . 31 4 _. 63.0 85 and e ........ . .'~-36.9 - - 895 Results o' calerolatton ot the abriCpee~lile tables for on mokersin Erre Counry.Pa 1or 1972-7A, by sex anm ape grou? 'Males I "- 1 .:1 3p-3' . 1 ~ 2 35-39 ........... r.3 404, . 3 .. , : a6-49 ..'-' t1 5034 13 ".85-5 ': .".'.'. . '30 - .. i 60-6c ......:..:....... ....... 40 . . i 65-69 ....... ....... :..... 41 . . . , . _. .. . ~ )0-14.:.:.:__-.....:..-' 39 1 ]6-29 ..:.':-...::.','.:.::L..... 69 I 80.a4 ............. .c....:103~ . 85 an0 al0er ............ ~. 76 '~ ` - Females. _ ~ ~ ~ I, 30-34 . ..,....... „ 5, 'I 35.39 .......: ................. 6~ ..:~ 40-44 ._..................... 14 I~ 45-<9 ..,:.:..:': ........... 22 50.5< . .. ..... w ' ..... ...... ~ 55.59 ................. .-.. 50 I 60.64 92 I 65-69 .............. ... ......... ln 0~I< .. ... 186 ~ )549 ..................... 328~, . :'-BD-B4 ................ ..1L.. 411 ~ 85 anC older ...:.......:. 466 6241` .00024 5 " U0120i 100009 62t5 .00o4g .54 -,00240 99.6BD 5,190 .00051 Sa A0255 99,64C 5,5e5 ,00196 .54' .00986 99,38E 6.983'AO18E. .53'.00926 -981606 5,431 00552 .52 .02724. 97,495 3,5)6.01116'.' .52' .05444. 94939 3,121 .0131: 52- . .06354'. 99 67E 2:646. .01472 .51 .07104 Ba,969 2A30 033co '-511 .15688 7e034 1,051', .098DC .46 .39050 65,767 1,046' q7c42 ... 1ADOC0 40,0e 5 120 499,]Il0 = 240 495,809 254 497,565 9BC 494AB0 91n 489,752 2.656 480!B35 5,163 461.287 5.70)' 434,112 5,965, 404,932 12.237 359,42> 25.682 254,63C 40,085200525 13.975 0015 .52' ,0C175 1DO.0C0 175 49Q562 11,969 OOD50 5< nc250 99ffi5 250 499.50D /3.641 .00102.54 .00509 99.575 507 495657 15.275 AOt4c .54 00119'99AEE. 5'11 493.562 14855 .00265 .53 .01337 9°]A 1.315 468A97 14,25] .00359 .52 Ot 735 9r 0c2 1,69c 491.000 11,5'6 -,00)94+ .52.03595 95,358 3;r15 46),502 9,295 .01518 ,52.0732391,643 6]11 441<37 8.556 .C219E .51 tn<tc 9<532 6.6<9 4n2.53? 7,3<2 .04c67.51 .20132 76 083 15.317. 342,122 4,831 .0853i.48. .34E31 60)66 21,185 250.91T 4:462 .1039)... 100000 39,6:1 39.601 198.1105 5A85945 50.9 4.569.2e5 Cfi.o 4,D81,445 40.9 3.5E5,BfiD 362 3.095400 31.5 2.605.64926] 2.124.81322.4 . 1,663.5261B.E 1i229,414 14.6 824,aB210.6 . 465.055 7.1 200,425 ' 5.C ~5,050,2e8 50.6 e 568 696 45?. 4;062.186 40.'9 J.565.5]9360 . 3.072,017 31.2 2.5s352026.6 2,102.523 22.0 1:65 016 17.8 1,150,5e1 14.1 791,044 10.4 446.922 7.4 198.005 ' 5.0 •Csnm.,.E. C.a.4 m us. ooe.~.uee ben u i aun.u er vr,n s,.u•uu. nnrE.l..d.n 1...r,nCle•4 cnnne M[Gy4n.J. He >uerle H.enx aqente

C Hammond, E- C.: Smoking in relation to the deathh rates of one million men and women.. in Epidemio- logieal approaches to the studyy of cancer and other chronicc diseases, edited by. W. Haensrll. 1 Natl Cancer Inso Monogr 19; 127-304 (1966). 24:. U.S. Bureau of the Census: Current housing reports. Series H-150-79. General housing eharaneristia for the United States and regions: 1979; Annual housing survey: 1979, p:.A. U-S Government Printing Otfice, Washington, D.C., 1981, Departmenn of Health, Education, and Welfare: Smok- ing and health-a report of the Surgeon General. DHEW Publication Nc. (PHS). 79-50066. U.S. Gov- ernment Printing Office, Washington,D.C., 1979. Chiang, C. L: lntroduction to stochastic ptocesses in, biostatistics. John Wiley & Sons, Inc., New York, 1968, Shryoak- H. S., and Siegel, J. S.: The methods an&rnn- lerials of demogrzphy. Third~ printing (revised). Us. Government Printing Office. W ashington, D.C., 1975. Keyf4tz N.: Introductionn to the mathematics of de- mography:.Addison-W'esley. Publishing Company, Read- ing. Mase., 1967. Lewis. C. E., and Lewis, M. A.: The potential impact 04 sexual equality on health. N Engl 3 Med 297: 863- 869 (1970). U.S. Department of Health and Human Services: The healthh consequences, of smoking for women-a report of the Surgeon General. U.S. Government Prinling. Office. Washington, D.C.. 1980. 16.Haensni, W., Loveland. D. B:, and Sirken, M.. G.: 23. Lung cancer monalityy as mlated'to residence and smoking historiess in whiR males. I Natl Cancer Inst 28: 947-1101 (1962)- 17. Godley, F. H.: Cigarette smoking,, social factors, and; 26. mortality: new estimates from national samples. Doc- loral dissertation, University of Maryland,. College 27. Park. 1974. t8'. Enstrom, J:. E., and Godley, F. H.:. Cancer mortality, among a reprr.entative sample of non-smokers in the 28. United States ducing 1966-1968. J Natl Cancer Inst 65c 1175-1183 (1980)~ 19. Casey, A. E., and Casey. 3. G.: Long-lived male popu- 24. lation. with hig?t, cholesterol intake, in Siieve Loughner, -. Ireland.. Ala J Med Sci 7: 21-26(1971 ). 20. Miller, G. H.: Malk-female longevity eomparisons 30. among the Amish. J Ind State Mcd Assoc 73: 471- 473 (1980). 21. Miller, G. H.: Smoking and longevity. J. 111 Lung.Assoc --- 39: 2-14 (1i976). 22. RoBot. E-, and Reid. D..D.: The validity of data from next-of-kin in studies of mortality among migrants. Int J Epidemiol.4: 51-54 (1975J. 23. U.S. Bureau of the Census: 1970census of housing. Housing characteristics for States. eities;, and counties.. Pt. 1. United States summary. U.S. Government Print-ing Office, Washington, D.C., 1972, vol. 1, p. 1-81. I Status of Nutrition Surveillance ctivities in 24 State and Melte{Zolitan Health Departments JAMES C. SCHE MS,, RD LAURA S.. SIMS. PhD, PH Mr. Scheer was formerlywith the Di iion of Nutrition- WIC Services. Tcnnesce Department of R lic Health. He is a nutrition consultant with Natural'Weight tro1, Inc., Nashvillc. Dr. Sims is. anassociate professor in the Nutrition , o- gram. College of Human Development. Pennsylvania Sta1e University, University Park., Teanheet requests toJames C. Scheer, MS„ RD. RD =3', Box 302. Nashvillc, Tenn, 37218. ' . SYNOPSIS ......................... . ..... ~ A.ctudy wa.rtotdertaken to examine nutrition sur- ~veillance activities and their usefulness in managing ;; programs of nutrition intervention. Questionnaires were returned by 24 of 26 diiectori of' nutrition units in State or metropolitan ftealtit departments participating in 1981 in the coordfnatednutrition surveillance system of the Centers for Disease Con- trol, which monitors high risk pediatric patients and pregnant women. The mean years of experience in surveillance ac- tivitks among the agencies was4, Only 25 percent o(the responding departments reported a self-suffi- cient computerized surveillance system. Personnel most involved in the coordinating; analyzing, and in- terpreting of the data were nutritionists who spent \erage of 17 hours per month. 03•734+703 Major u o! surveillance data reported for ptu- poses of the nut 'on programs were to (a) identify collection sites with .b[enu such aserrors in mea- surirrg heights andweight d hematocrits warrant- ing checks~ for quality cont.ol, dejtne thee extent of nutrition-related disortiens in t target popula- tions,. (.e) provide objective focaf dat m assist in Jury-J•roca tan, Vel. H. RC- 4 341

. H. ..standardceruusdata'ondiTe--;- -- expectancy in tlae'Uttited States and:- :d elsewPure merg'e the nvo,verydifjerenr ~ _ morfaftry rates of s»iokersand I . nonsmokns, producing iitaccurate ~- '-estimatca for each caregory whenthese- :. eategoriesareconsideredseparatelyt ,- . Therefore,., in any discussion on ¢ch~aeiai - "useofthemale-felelogFriry." man , di$erenec base[lansuehmerged dafc,'- , , -for esarnpte, calculation~o/ pensior ~. -'t -`'•benefttsorfifeinsurancepremuM :rates, - th a ' ese c tegorical difjereriee shi .,euce°. j b ' aakenintoaccount-"~`_ become similar to male behavior, and they aeom- mented on the adverse impact of this~ tread on fe- male health and morbidity. The most signifirmat change fn~ younger women's healttihabits in the United States.over the past.deeade has been a large increase in teenage and preteenagee smoking. Tbe 1981'. Surgeon GeneraSs report on t&health coo- sequencesof smoking forwomen.(3D)'n ihdicates that teenage girls have suroassad4eenage boys Vn the per- centage of smokers, Whean cobor!s of women wito have smoked asmuehas men reach the later decades of life. the results of our. stud}° suggest that their livess will be shortened as much as mea's and that the presen!, diflerer•.cesin longevitym between men and women ('MFLD) wilPdisappear, - Reierences .' ................. . 1, 115. Bv e u oftn Ce ws: Stavitial ebatract of the III Lt d Siates. US..Go ernmanl P intingofice, Waah. gion D C., 1999 p)0. 2 F'sner, R. A.: Th canuen cvnrruversy:.Ollver end` relatives, who may not have been familiar with the Boyd, Ltd., emnburpai _toe9: ~early: Iife smoking habits of the deccdenrs, Also, ~Berkson.l.: Thee stavistical nuny of assodmlun bet.een ~~ smukineand rebacco ProceeAings of SisH MeninRa of manv respondent> have a tendency to aulomaacally' . classif: themselves or anyone who has quit smoking S19Maaf`.Chnr' Roch•sie- Mmn. 1955 vol, 3q Pp. Y .~ at some time during their lives as nonsmokers even G Montagv, A.:The mraral xuperibntv ol women. Tne _. i when they. knorrv of past smoking. This. imprecision Macmillan Compam. NeoYork:. !9s3, y. 203. I/) results in the elassifieelion of deceased fbrlner smok- 5. {•ondon. W. T., Resenoerg.. 5. E,, Draper;,l. W., and ers as nonsmokers and thusinereasess the mortality . Almy, T. P.: The efiauof esuogens on ethcrosc!croess. 63'-69 U961). rales.attribuled to nonsmokers. In our srudy,.we were Ann lmcrn hfed 55: 6. Swmlr•. 1~.. er at.: ERCaiveneca ob esrogcns Soe t6c able to minimize this.potenllallvs confoundingfaerorg lunpeerm iherapv oCmiddle.aged inen witb,e hlnury oflnvestigatorsconducting new studies should use the myocardis! inrarNCr,. In Coronary hem dixues, .,. mostl preciset method'sfor elaSsifiCation of smaking edrted by W. Liknfi and 1. H. Muye[. The SevmttiStatus abrremann SSSmposium, Grunt & Sttanoo, Ner YorII.. b . H a rremann-19fi3, pp, a16-<35. 'plie results of our studpy have two, implications. r:'Rnaenman,.R..H., ct all: A pndictive study of eonr FirSt, standard censuss data oali(e expeulanc;~ ih the nary hearr d/sease. TAe. Western Collzbarativc Group UnitechStatCs and elsewhere merge the two very difb Smdy.JAMA I89: I5-]6 (1966). ferelnt', moitalf[y ratesof snloker5and nonsmokCrf, a. Russek, H, L: Stress, tobacci+ and cmanary dlsease producinginaccura®g estimale> forr each category in Nurth Amuiczr Grofarlorul eroups. IAMAt9:: B9-94 (1965). when these categories arc ennsidetEd separately: p]enkirv C. D.: Paycholagicxl and soclsl Therefore, in anv discussion or actuarial use of the eromrsors of . orunaryy di'vcaae. Nr. 11 F Enpl 1 Mea 294 : 24L253 . MFLD based on such, merged data, fbr example, (t91, 1). -1 caleulationrof'penslon benefltsor IifE Insuran¢e- pre- 10 . Jenkim. C. D-: PxycholoRical'and socixl pre.vreucy oG . Imium rates;..lhae categorical diHerences should be ' cornnarv.diseana.,Pi :: N Ensl I Mcd 28<: 301-3Pf' I take. _. .. . , (19]1) .. L taken mto aeeounL Jt Hayes 5. G., a d E nleib, M women, work aod The second'~imphcallon is.that a p rson's sex-role -y hearn a a, : P p n ve hndmsa trom the behavior has an ifnpacl on heallh andmorbidih and, F rs m Nc I Siudy. AI Public Hcalth R0: 1 oonsequently; on long in'. Ima reccnt arricle, Lrwis1+3 lel /'!9F0) I and Le wis(19)attribu.ed the present. MFLD to I 11. W Idro , L: W'hyy d vmer hve longer Iham men' I mensexcessive xmokir! , Iheir suicidal and accideno- 5 S Md !n 3 s a(tsvs) .. I risk Fehavior: Ihein greater aleoRol consumption, and'i ' 13. Pmsmn. S. H.- Aa imernsuona'. comparia,not nxcn- s,ve advlr.monzlicy: PcnS:eE 24: 5-20 (19]Ol. , i m6lc dlsdain for •'minOr' medie2l aSiistanee, These ~IL. Rcrncrfort, re. D.: Toh.reo smoking and the sc mot- ., authors also noted the trends for femalee behavior tonhndinercnrral. Demoprophv 9: 203-216(1991). va .eerne u..ue nav.na